DOCSLIB.ORG

Edger: Differential Expression Analysis of Digital Gene

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Edger: Differential Expression Analysis of Digital Gene edgeR: differential expression analysis of digital gene expression data User's Guide Yunshun Chen, Davis McCarthy, Matthew Ritchie, Mark Robinson, Gordon K. Smyth First edition 17 September 2008 Last revised 30 June 2016 Contents 1 Introduction 5 1.1 Scope . 5 1.2 Citation . 6 1.3 How to get help . 7 1.4 Quick start . 8 2 Overview of capabilities 9 2.1 Terminology . 9 2.2 Aligning reads to a genome . 9 2.3 Producing a table of read counts . 9 2.4 Reading the counts from a file . 10 2.5 The DGEList data class . 10 2.6 Filtering . 11 2.7 Normalization . 12 2.7.1 Normalization is only necessary for sample-specific effects . 12 2.7.2 Sequencing depth . 12 2.7.3 RNA composition . 12 2.7.4 GC content . 13 2.7.5 Gene length . 13 2.7.6 Model-based normalization, not transformation . 13 2.7.7 Pseudo-counts . 14 2.8 Negative binomial models . 14 2.8.1 Introduction . 14 2.8.2 Biological coefficient of variation (BCV) . 15 2.8.3 Estimating BCVs . 16 2.8.4 Quasi negative binomial . 17 2.9 Pairwise comparisons between two or more groups (classic) . 17 2.9.1 Estimating dispersions . 17 2.9.2 Testing for DE genes . 18 2.10 More complex experiments (glm functionality) . 19 2.10.1 Generalized linear models . 19 1 2.10.2 Estimating dispersions . 19 2.10.3 Testing for DE genes . 20 2.11 What to do if you have no replicates . 21 2.12 Differential expression above a fold-change threshold . 23 2.13 Gene ontology (GO) and pathway analysis . 24 2.14 Gene set testing . 25 2.15 Clustering, heatmaps etc . 25 2.16 Alternative splicing . 26 2.17 CRISPR-Cas9 and shRNA-seq screen analysis . 26 3 Specific experimental designs 28 3.1 Introduction . 28 3.2 Two or more groups . 28 3.2.1 Introduction . 28 3.2.2 Classic approach . 29 3.2.3 GLM approach . 30 3.2.4 Questions and contrasts . 31 3.2.5 A more traditional glm approach . 32 3.2.6 An ANOVA-like test for any differences . 33 3.3 Experiments with all combinations of multiple factors . 34 3.3.1 Defining each treatment combination as a group . 34 3.3.2 Nested interaction formulas . 36 3.3.3 Treatment effects over all times . 36 3.3.4 Interaction at any time . 37 3.4 Additive models and blocking . 37 3.4.1 Paired samples . 37 3.4.2 Blocking . 38 3.4.3 Batch effects . 40 3.5 Comparisons both between and within subjects . 40 4 Case studies 43 4.1 RNA-Seq of oral carcinomas vs matched normal tissue . 43 4.1.1 Introduction . 43 4.1.2 Reading in the data . 43 4.1.3 Annotation . 44 4.1.4 Filtering and normalization . 45 4.1.5 Data exploration . 46 4.1.6 The design matrix . 46 4.1.7 Estimating the dispersion . 47 4.1.8 Differential expression . 48 4.1.9 Gene ontology analysis . 50 2 4.1.10 Setup . 51 4.2 RNA-Seq of pathogen inoculated arabidopsis with batch effects . 52 4.2.1 Introduction . 52 4.2.2 RNA samples . 52 4.2.3 Loading the data . 52 4.2.4 Filtering and normalization . 53 4.2.5 Data exploration . 53 4.2.6 The design matrix . 54 4.2.7 Estimating the dispersion . 55 4.2.8 Differential expression . 56 4.2.9 Setup . 58 4.3 Profiles of Yoruba HapMap individuals . 59 4.3.1 Background . 59 4.3.2 Loading the data . 59 4.3.3 Filtering and normalization . 61 4.3.4 Estimating the dispersion . 62 4.3.5 Differential expression . 63 4.3.6 Gene set testing . 64 4.3.7 Setup . 66 4.4 RNA-Seq profiles of mouse mammary gland . 67 4.4.1 Introduction . 67 4.4.2 Read alignment and processing . 67 4.4.3 Count loading and annotation . 68 4.4.4 Filtering and normalization . 69 4.4.5 Data exploration . 70 4.4.6 The design matrix . 71 4.4.7 Estimating the dispersion . 72 4.4.8 Differential expression . 74 4.4.9 ANOVA-like testing . 76 4.4.10 Gene ontology analysis . 77 4.4.11 Gene set testing . 79 4.4.12 Setup . 80 4.5 Differential splicing after Pasilla knockdown . 81 4.5.1 Introduction . 81 4.5.2 RNA-Seq samples . 81 4.5.3 Read alignment and processing . 82 4.5.4 Count loading and annotation . 83 4.5.5 Filtering and normalization . 84 4.5.6 Data exploration . 85 4.5.7 The design matrix . 86 4.5.8 Estimating the dispersion . 86 3 4.5.9 Differential expression . 88 4.5.10 Alternative splicing . 88 4.5.11 Setup . 91 4.5.12 Acknowledgements . 91 4.6 CRISPR-Cas9 knockout screen analysis . 91 4.6.1 Introduction . 91 4.6.2 Sequence processing . ..
Load more

Recommended publications
  • The Transition from Primary Colorectal Cancer to Isolated Peritoneal Malignancy
    medRxiv preprint doi: https://doi.org/10.1101/2020.02.24.20027318; this version posted February 25, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license . The transition from primary colorectal cancer to isolated peritoneal malignancy is associated with a hypermutant, hypermethylated state Sally Hallam1, Joanne Stockton1, Claire Bryer1, Celina Whalley1, Valerie Pestinger1, Haney Youssef1, Andrew D Beggs1 1 = Surgical Research Laboratory, Institute of Cancer & Genomic Science, University of Birmingham, B15 2TT. Correspondence to: Andrew Beggs, [email protected] KEYWORDS: Colorectal cancer, peritoneal metastasis ABBREVIATIONS: Colorectal cancer (CRC), Colorectal peritoneal metastasis (CPM), Cytoreductive surgery and heated intraperitoneal chemotherapy (CRS & HIPEC), Disease free survival (DFS), Differentially methylated regions (DMR), Overall survival (OS), TableFormalin fixed paraffin embedded (FFPE), Hepatocellular carcinoma (HCC) ARTICLE CATEGORY: Research article NOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice. 1 medRxiv preprint doi: https://doi.org/10.1101/2020.02.24.20027318; this version posted February 25, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY 4.0 International license . NOVELTY AND IMPACT: Colorectal peritoneal metastasis (CPM) are associated with limited and variable survival despite patient selection using known prognostic factors and optimal currently available treatments.
    [Show full text]
  • Sustained Intrinsic WNT and BMP4 Activation Impairs Hesc Diferentiation to Defnitive Endoderm and Drives the Cells Towards Extra‑Embryonic Mesoderm C
    www.nature.com/scientificreports OPEN Sustained intrinsic WNT and BMP4 activation impairs hESC diferentiation to defnitive endoderm and drives the cells towards extra‑embryonic mesoderm C. Markouli1,3, E. Couvreu De Deckersberg1,3, D. Dziedzicka1, M. Regin1, S. Franck1, A. Keller1, A. Gheldof2, M. Geens1, K. Sermon1 & C. Spits1* We identifed a human embryonic stem cell subline that fails to respond to the diferentiation cues needed to obtain endoderm derivatives, diferentiating instead into extra‑embryonic mesoderm. RNA‑sequencing analysis showed that the subline has hyperactivation of the WNT and BMP4 signalling. Modulation of these pathways with small molecules confrmed them as the cause of the diferentiation impairment. While activation of WNT and BMP4 in control cells resulted in a loss of endoderm diferentiation and induction of extra‑embryonic mesoderm markers, inhibition of these pathways in the subline restored its ability to diferentiate. Karyotyping and exome sequencing analysis did not identify any changes in the genome that could account for the pathway deregulation. These fndings add to the increasing evidence that diferent responses of stem cell lines to diferentiation protocols are based on genetic and epigenetic factors, inherent to the line or acquired during cell culture. Human embryonic stem cells (hESCs) are a potent tool for the study of early development, in disease modelling and in regenerative medicine, as they can diferentiate to any cell type of the human body. HESC diferentiation protocols aim to mimic the fne orchestration of time dependent pathway modulation which is observed in vivo. Given this complexity, it is broadly acknowledged that individual cell lines may display a preference for difer- entiation to one germ layer over another1,2.
    [Show full text]
  • Peripheral Nerve Single-Cell Analysis Identifies Mesenchymal Ligands That Promote Axonal Growth
    Research Article: New Research Development Peripheral Nerve Single-Cell Analysis Identifies Mesenchymal Ligands that Promote Axonal Growth Jeremy S. Toma,1 Konstantina Karamboulas,1,ª Matthew J. Carr,1,2,ª Adelaida Kolaj,1,3 Scott A. Yuzwa,1 Neemat Mahmud,1,3 Mekayla A. Storer,1 David R. Kaplan,1,2,4 and Freda D. Miller1,2,3,4 https://doi.org/10.1523/ENEURO.0066-20.2020 1Program in Neurosciences and Mental Health, Hospital for Sick Children, 555 University Avenue, Toronto, Ontario M5G 1X8, Canada, 2Institute of Medical Sciences University of Toronto, Toronto, Ontario M5G 1A8, Canada, 3Department of Physiology, University of Toronto, Toronto, Ontario M5G 1A8, Canada, and 4Department of Molecular Genetics, University of Toronto, Toronto, Ontario M5G 1A8, Canada Abstract Peripheral nerves provide a supportive growth environment for developing and regenerating axons and are es- sential for maintenance and repair of many non-neural tissues. This capacity has largely been ascribed to paracrine factors secreted by nerve-resident Schwann cells. Here, we used single-cell transcriptional profiling to identify ligands made by different injured rodent nerve cell types and have combined this with cell-surface mass spectrometry to computationally model potential paracrine interactions with peripheral neurons. These analyses show that peripheral nerves make many ligands predicted to act on peripheral and CNS neurons, in- cluding known and previously uncharacterized ligands. While Schwann cells are an important ligand source within injured nerves, more than half of the predicted ligands are made by nerve-resident mesenchymal cells, including the endoneurial cells most closely associated with peripheral axons. At least three of these mesen- chymal ligands, ANGPT1, CCL11, and VEGFC, promote growth when locally applied on sympathetic axons.
    [Show full text]
  • Supplementary Material Study Sample the Vervets Used in This
    Supplementary Material Study sample The vervets used in this study are part of a pedigreed research colony that has included more than 2,000 monkeys since its founding. Briefly, the Vervet Research Colony (VRC) was established at UCLA during the 1970’s and 1980’s from 57 founder animals captured from wild populations on the adjacent Caribbean islands of St. Kitts and Nevis; Europeans brought the founders of these populations to the Caribbean from West Africa in the 17th Century 1. The breeding strategy of the VRC has emphasized the promotion of diversity, the preservation of the founding matrilines, and providing all females and most of the males an opportunity to breed. The colony design modeled natural vervet social groups to facilitate behavioral investigations in semi-natural conditions. Social groups were housed in large outdoor enclosures with adjacent indoor shelters. Each enclosure had chain link siding that provided visual access to the outside, with one or two large sitting platforms and numerous shelves, climbing structures and enrichments devices. The monkeys studied were members of 16 different social matrilineal groups, containing from 15 to 46 members per group. In 2008 the VRC was moved to Wake Forest School of Medicine’s Center for Comparative Medicine Research, however the samples for gene expression measurements in Dataset 1 (see below) and the MRI assessments used in this study occurred when the colony was at UCLA. Gene expression phenotypes Two sets of gene expression measurements were collected. Dataset 1 used RNA obtained from whole blood in 347 vervets, assayed by microarray (Illumina HumanRef-8 v2); Dataset 2 assayed gene expression by RNA-Seq, in RNA obtained from 58 animals, with seven tissues (adrenal, blood, Brodmann area 46 [BA46], caudate, fibroblast, hippocampus and pituitary) measured in each animal.
    [Show full text]
  • Downloaded from the European Nucleotide Archive (ENA; 126
    Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 5 September 2018 doi:10.20944/preprints201809.0082.v1 1 Article 2 Transcriptomics as precision medicine to classify in 3 vivo models of dietary-induced atherosclerosis at 4 cellular and molecular levels 5 Alexei Evsikov 1,2, Caralina Marín de Evsikova 1,2* 6 1 Epigenetics & Functional Genomics Laboratory, Department of Molecular Medicine, Morsani College of 7 Medicine, University of South Florida, Tampa, Florida, 33612, USA; 8 2 Department of Research and Development, Bay Pines Veteran Administration Healthcare System, Bay 9 Pines, FL 33744, USA 10 11 * Correspondence: [email protected]; Tel.: +1-813-974-2248 12 13 Abstract: The central promise of personalized medicine is individualized treatments that target 14 molecular mechanisms underlying the physiological changes and symptoms arising from disease. 15 We demonstrate a bioinformatics analysis pipeline as a proof-of-principle to test the feasibility and 16 practicality of comparative transcriptomics to classify two of the most popular in vivo diet-induced 17 models of coronary atherosclerosis, apolipoprotein E null mice and New Zealand White rabbits. 18 Transcriptomics analyses indicate the two models extensively share dysregulated genes albeit with 19 some unique pathways. For instance, while both models have alterations in the mitochondrion, the 20 biochemical pathway analysis revealed, Complex IV in the electron transfer chain is higher in mice, 21 whereas the rest of the electron transfer chain components are higher in the rabbits. Several fatty 22 acids anabolic pathways are expressed higher in mice, whereas fatty acids and lipids degradation 23 pathways are higher in rabbits.
    [Show full text]
  • Clustering & Function Enrichment Analysis
    Clustering & Function Enrichment Analysis Qi Sun Bioinformatics Facility Cornell University Clustering analysis 1.Hierarchical 2.K-means 3.Co-expression network Prepare data for clustering Step 1. LOG transformation of CPM value to improve the distribution 7000 3500 6000 3000 5000 2500 4000 2000 3000 1500 2000 1000 1000 500 0 0 4 8 12 16 20 24 28 32 36 40 44 48 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 CPM Log2(CPM) Step 2. Remove genes with no variation across samples Clustering analysis on multiple conditions of RNA-seq data 1 2 3 4 Hierarchical clustering K-means clustering $TRINITY_HOME/Analysis/DifferentialExpression/ define_clusters_by_cutting_tree.pl -R diffExpr.P0.001_C2.matrix.RData -K 18 K-means clustering K value need to be pre-specified Co-expression network modules WGCNA (weighted correlation network analysis) • transform the initial distance matrix into Topological Overlap Matrix http://rgm3.lab.nig.ac.jp/RGM/R_image_list?package=WGCNA&init=true Connect RNA-seq results to biological pathways RNA-seq results 169 genes up- regulated 340 genes down- regulated Mapman pathway viewer https://doi.org/10.1111/j.1365-3040.2009.01978.x From a biological question to a statistical problem In the whole genome RNA-seq results Number of genes: 29960 Number of DE genes: 297 Genes in P53 pathway: 40 DE genes in P53: 3 If the P53 genes over-represented in the DE genes? How to tag a gene with function/pathway categories? • Free text description Gene ID Gene description GRMZM2G002950 Putative leucine-rich repeat receptor-like protein kinase
    [Show full text]
  • Edger: Differential Analysis of Sequence Read Count Data User's
    edgeR: differential analysis of sequence read count data User’s Guide Yunshun Chen 1,2, Davis McCarthy 3,4, Matthew Ritchie 1,2, Mark Robinson 5, and Gordon Smyth 1,6 1Walter and Eliza Hall Institute of Medical Research, Parkville, Victoria, Australia 2Department of Medical Biology, University of Melbourne, Victoria, Australia 3St Vincent’s Institute of Medical Research, Fitzroy, Victoria, Australia 4Melbourne Integrative Genomics, University of Melbourne, Victoria, Australia 5Institute of Molecular Life Sciences and SIB Swiss Institute of Bioinformatics, University of Zurich, Zurich, Switzerland 6School of Mathematics and Statistics, University of Melbourne, Victoria, Australia First edition 17 September 2008 Last revised 12 May 2021 Contents 1 Introduction .............................7 1.1 Scope .................................7 1.2 Citation.................................7 1.3 How to get help ............................9 1.4 Quick start ............................... 10 2 Overview of capabilities ...................... 11 2.1 Terminology .............................. 11 2.2 Aligning reads to a genome ..................... 11 2.3 Producing a table of read counts .................. 11 2.4 Reading the counts from a file ................... 12 2.5 Pseudoalignment and quasi-mapping ............... 12 2.6 The DGEList data class ....................... 12 2.7 Filtering ................................ 13 2.8 Normalization ............................. 14 2.8.1 Normalization is only necessary for sample-specific effects .... 14 2.8.2 Sequencing depth ........................ 14 2.8.3 Effective library sizes ....................... 15 2.8.4 GC content ............................ 15 2.8.5 Gene length ........................... 16 2.8.6 Model-based normalization, not transformation .......... 16 2.8.7 Pseudo-counts .......................... 16 2 edgeR User’s Guide 2.9 Negative binomial models ...................... 17 2.9.1 Introduction ........................... 17 2.9.2 Biological coefficient of variation (BCV) .............
    [Show full text]
  • Sulfotransferase 1A3 (SULT1A3)
    A Dissertation entitled Functional Genomic Studies On The Genetic Polymorphisms Of The Human Cytosolic Sulfotransferase 1A3 (SULT1A3) by Ahsan Falah Hasan Bairam Submitted to the Graduate Faculty as partial fulfillment of the requirements for the Doctor of Philosophy Degree in Experimental Therapeutics ________________________________________ Dr. Ming-Cheh Liu, Committee Chair ________________________________________ Dr. Ezdihar Hassoun, Committee Member ________________________________________ Dr. Zahoor Shah, Committee Member ________________________________________ Dr. Caren Steinmiller, Committee Member ________________________________________ Dr. Amanda Bryant-Friedrich, Dean College of Graduate Studies The University of Toledo May 2018 Copyright 2018, Ahsan Falah Hasan Bairam This document is copyrighted material. Under copyright law, no parts of this document may be reproduced without the expressed permission of the author. An Abstract of Functional Genomic Studies On The Genetic Polymorphisms Of The Human Cytosolic Sulfotransferase 1A3 (SULT1A3) by Ahsan Falah Hasan Bairam Submitted to the Graduate Faculty as partial fulfillment of the requirements for the Doctor of Philosophy Degree in Experimental Therapeutics (Pharmacology/Toxicology) The University of Toledo May 2018 Abstract Previous studies have demonstrated the involvement of sulfoconjugation in the metabolism of catecholamines and serotonin (5-HT), as well as a wide range of xenobiotics including drugs. The study presented in this dissertation aimed to clarify the effects of coding single nucleotide polymorphisms (cSNPs) of the human SULT1A3 and SULT1A4 genes on the enzymatic characteristics of the sulfation of catecholamines, 5- HT, and selected drugs by SULT1A3 allozymes. Following a comprehensive search of different SULT1A3 and SULT1A4 genotypes, thirteen non-synonymous (missense) cSNPs of SULT1A3/SULT1A4 were identified. cDNAs encoding the corresponding SULT1A3 allozymes, packaged in pGEX-2T vector were generated by site-directed mutagenesis.
    [Show full text]
  • De Novo Transcriptome Assembly, Functional Annotation, And
    www.nature.com/scientificreports OPEN De novo transcriptome assembly, functional annotation, and expression profling of rye (Secale cereale L.) hybrids inoculated with ergot (Claviceps purpurea) Khalid Mahmood1,2*, Jihad Orabi1, Peter Skov Kristensen1, Pernille Sarup1, Lise Nistrup Jørgensen2 & Ahmed Jahoor1,3 Rye is used as food, feed, and for bioenergy production and remain an essential grain crop for cool temperate zones in marginal soils. Ergot is known to cause severe problems in cross-pollinated rye by contamination of harvested grains. The molecular response of the underlying mechanisms of this disease is still poorly understood due to the complex infection pattern. RNA sequencing can provide astonishing details about the transcriptional landscape, hence we employed a transcriptomic approach to identify genes in the underlying mechanism of ergot infection in rye. In this study, we generated de novo assemblies from twelve biological samples of two rye hybrids with identifed contrasting phenotypic responses to ergot infection. The fnal transcriptome of ergot susceptible (DH372) and moderately ergot resistant (Helltop) hybrids contain 208,690 and 192,116 contigs, respectively. By applying the BUSCO pipeline, we confrmed that these transcriptome assemblies contain more than 90% of gene representation of the available orthologue groups at Virdiplantae odb10. We employed a de novo assembled and the draft reference genome of rye to count the diferentially expressed genes (DEGs) between the two hybrids with and without inoculation. The gene expression comparisons revealed that 228 genes were linked to ergot infection in both hybrids. The genome ontology enrichment analysis of DEGs associated them with metabolic processes, hydrolase activity, pectinesterase activity, cell wall modifcation, pollen development and pollen wall assembly.
    [Show full text]
  • Co-Expression of Mklf7 and Trka 1149
    Development 128, 1147-1158 (2001) 1147 Printed in Great Britain © The Company of Biologists Limited 2001 DEV4514 mKlf7, a potential transcriptional regulator of TrkA nerve growth factor receptor expression in sensory and sympathetic neurons Lei Lei, Long Ma, Serge Nef, To Thai and Luis F. Parada* Center for Developmental Biology and Kent Waldrep Foundation Center for Basic Research on Nerve Growth and Regeneration, University of Texas, Southwestern Medical Center, 6000 Harry Hines Boulevard, Dallas, TX 75390-9133, USA *Author for correspondence (e-mail: [email protected]) Accepted 17 January; published on WWW 13 March 2001 SUMMARY Development of the nervous system relies on stringent Kruppel-like factor, named mKlf7, binds specifically to an regulation of genes that are crucial to this process. TrkA, Ikaros core binding element that is crucial for in vivo the receptor for nerve growth factor (NGF), is tightly TrkA enhancer function. Using in situ hybridization, we regulated during embryonic development and is essential demonstrate that mKlf7 is coexpressed with TrkA in for the survival and differentiation of neural crest-derived sensory and sympathetic neurons during embryogenesis sensory and sympathetic neurons. We have previously and in adulthood. These data are consistent with the idea identified a mouse TrkA enhancer and have characterized that mKlf7 may directly regulate TrkA gene expression in several cis regulatory elements that are important for the peripheral nervous system. appropriate TrkA expression in vivo. We now report the cloning of a novel gene encoding a Kruppel-like factor from Key words: Enhancer, Kruppel-like factor, Neurotrophin, Neural a mouse dorsal root ganglion expression library.
    [Show full text]
  • SUPPLEMENTARY MATERIALS Figure S1. Pre-Processing
    SUPPLEMENTARY MATERIALS Figure S1. Pre-processing the raw data of the pancreatic circulating tumor cells (CTCs). (A) Total read counts of single-cell RNA sequencing data of 75 murine pancreatic CTCs. The red horizontal line represents the cutoff, two counts per million (CPM). Three samples below this number were excluded. (B) Distribution of transformed data using a boxplot. The figure is generated by iDEP 9.0 (http://bioinformatics.sdstate.edu/idep/). Figure S2: The classification of 72 samples. (A) Principal component analysis (PCA) scores plot of 72 samples. In this figure, x and y axes represent the principal component 1 (PC1) = 16.6% variance and PC2 = 6.1% variance, respectively. All samples could be divided into three distinct clusters. The corresponding ellipses were plotted based on a 95% probability from the same group. The figure is generated by ClustVis (https://biit.cs.ut.ee/clustvis/). (B) The heatmap of 3000 genes expression. The relative color scheme was established, in which the maximum values were converted to red and the minimum convert to green in each row. We used the Pearson correlation as a metric to do the hierarchical clustering with average linkage in both columns and rows. The visualization was done by the MORPHEUS (https://software.broadinstitute.org/morpheus/). Figure S3: The differentially expressed genes. (A) The volcano plot of differentially expressed genes. Genes with |Fold change (FC) | > 2 and a corrected p-value, false discovery rate (FDR) < 0.050 (Benjamini- Hochberg procedure) were defined as differentially expressed genes. Red dots represent up-regulated genes and blue dots down-regulated genes.
    [Show full text]
  • Data-Driven and Knowledge-Driven Computational Models of Angiogenesis in Application to Peripheral Arterial Disease
    DATA-DRIVEN AND KNOWLEDGE-DRIVEN COMPUTATIONAL MODELS OF ANGIOGENESIS IN APPLICATION TO PERIPHERAL ARTERIAL DISEASE by Liang-Hui Chu A dissertation submitted to Johns Hopkins University in conformity with the requirements for the degree of Doctor of Philosophy Baltimore, Maryland March, 2015 © 2015 Liang-Hui Chu All Rights Reserved Abstract Angiogenesis, the formation of new blood vessels from pre-existing vessels, is involved in both physiological conditions (e.g. development, wound healing and exercise) and diseases (e.g. cancer, age-related macular degeneration, and ischemic diseases such as coronary artery disease and peripheral arterial disease). Peripheral arterial disease (PAD) affects approximately 8 to 12 million people in United States, especially those over the age of 50 and its prevalence is now comparable to that of coronary artery disease. To date, all clinical trials that includes stimulation of VEGF (vascular endothelial growth factor) and FGF (fibroblast growth factor) have failed. There is an unmet need to find novel genes and drug targets and predict potential therapeutics in PAD. We use the data-driven bioinformatic approach to identify angiogenesis-associated genes and predict new targets and repositioned drugs in PAD. We also formulate a mechanistic three- compartment model that includes the anti-angiogenic isoform VEGF165b. The thesis can serve as a framework for computational and experimental validations of novel drug targets and drugs in PAD. ii Acknowledgements I appreciate my advisor Dr. Aleksander S. Popel to guide my PhD studies for the five years at Johns Hopkins University. I also appreciate several professors on my thesis committee, Dr. Joel S. Bader, Dr.
    [Show full text]